Pulmonary perfusion by chest digital dynamic radiography: Comparison between breath‐holding and deep‐breathing acquisition

Abstract Purpose Pulmonary perfusion is an important factor for gas exchange. Chest digital dynamic radiography (DDR) by the deep‐breathing protocol can evaluate pulmonary perfusion in healthy subjects. However, respiratory artifacts may affect DDR in patients with respiratory diseases. We examined the feasibility of a breath‐holding protocol and compared it with the deep‐breathing protocol to reduce respiratory artifacts. Materials and methods A total of 42 consecutive patients with respiratory diseases (32 males; age, 68.6 ± 12.3 yr), including 21 patients with chronic obstructive pulmonary disease, underwent chest DDR through the breath‐holding protocol and the deep‐breathing protocol. Imaging success rate and exposure to radiation were compared. The correlation rate of temporal changes in each pixel value between the lung fields and left cardiac ventricles was analyzed. Results Imaging success rate was higher with the breath‐holding protocol vs the deep‐breathing protocol (97% vs 69%, respectively; P < 0.0001). The entrance surface dose was lower with the breath‐holding protocol (1.09 ± 0.20 vs 1.81 ± 0.08 mGy, respectively; P < 0.0001). The correlation rate was higher with the breath‐holding protocol (right lung field, 41.7 ± 9.3%; left lung field, 44.2 ± 8.9% vs right lung field, 33.4 ± 6.6%; left lung field, 36.0 ± 7.1%, respectively; both lung fields, P < 0.0001). In the lower lung fields, the correlation rate was markedly different (right, 15.3% difference; left, 14.1% difference; both lung fields, P < 0.0001). Conclusion The breath‐holding protocol resulted in high imaging success rate among patients with respiratory diseases, yielding vivid images of pulmonary perfusion.


| INTRODUCTION
Pulmonary perfusion is the circulation of blood from the right ventricle to the left atrium and involved in gas exchange. 1 Therefore, the evaluation of pulmonary circulation is of vital importance for the management of numerous respiratory diseases. Pulmonary perfusion is generally evaluated based on examinations using contrast media or radioactive agents. 2,3 Computed tomography (CT) and magnetic resonance imaging (MRI) have been beneficial in assessing the physiological function of the lungs. 4,5 However, CT and MRI have disadvantages, including high doses of radiation in CT scanning, 6 long duration of examination, and substantial financial burden associated with MRI. 7 In these diagnostic imaging methods, the patients need to be in the supine position, and imaging in the upright position is challenging.
Chest digital dynamic radiography (DDR) has been developed for the novel evaluation of pulmonary perfusion. [8][9][10] Digital dynamic radiography technology rapidly captures sequential radiographs in a single examination to observe the dynamic interactions of anatomical structures. The lungs always hold approximately 500 ml of blood, and approximately 75 ml of blood is transferred into the vessels during a single cardiac cycle. The pulsatile blood flow is captured on the image as a change in the pixel value of the pulmonary vessels. Technically speaking, the pixel value increases very slightly during systole and decreases very slightly during diastole. Chest DDR can detect the change in pixel value synchronized with the cardiac cycle, and generate vivid pulmonary perfusion images. 8,11,12 Its duration is approximately 15 s, it does not require contrast media, and can be performed in the standing, seated, and supine positions. Owing to the high sensitivity of the flat panel detector, the total exposure to radiation can be less than the limits for two projections recommended by the International Atomic Energy Agency (1.9 mGy). 13 Based on these characteristics, chest DDR may be beneficial to screening for pulmonary embolism or pulmonary hypertension. 8 A previous animal-based study of a porcine pulmonary embolism model has revealed that chest DDR was able to detect perfusion defects in the lobe unit. 14 In clinical practice, it has been reported that chest DDR could detect pulmonary perfusion defects in patients with giant cell arteritis and chronic thromboembolic pulmonary hypertension. 15,16 Moreover, DDR images were remarkably similar to those of pulmonary angiography, the iodine mapping images acquired with dual-energy CT, and those of lung perfusion scintigraphy. 11,12 The deep-breathing protocol is an imaging protocol that has been used since the dawn of DDR technology. 8 It was an innovative method for the simultaneous evaluation of both ventilation and pulmonary blood flow by pausing for 2 s between deep breaths. However, respiratory artifacts are an important challenge for the examination of patients with respiratory diseases. Incomplete diaphragm pause in chest DDR could lead to two types of respiratory artifacts, namely blurred images and misalignment of the rib cage. In a previous study of chest DDR conducted by Tanaka et al., breathholding at rest expiratory level was helpful in capturing vivid pulmonary perfusion images. 12 Nevertheless, the appropriate duration of breath-holding or levels (e.g., maximum or rest, expiratory or inspiratory) are unclear. For clinical use, a shorter breath-holding time is preferable. Thus, we sought to investigate whether breath-holding for 6 s (breath-holding protocol) at the maximum inspiratory level could reduce respiratory artifacts in clinical practice.
For this purpose, we examined the feasibility of the breath-holding protocol and compared it with that of the deep-breathing protocol, as previously reported. 8 2 | MATERIALS AND METHODS

2.A | Patients
From February 2019 to January 2020, 44 consecutive patients with respiratory diseases who visited our hospital were investigated. The inclusion criteria for patients were: age >20 yr; chest DDR by both protocols; and a lung function test. Initially, 59 patients were photographed during the relevant period. The two protocols were utilized in 49 patients. After excluding five patients who did not undergo a lung function test and two patients with left bundle block, 42 patients were included in the analysis (Fig. 1). Given the principle of imaging, the reliability of pulmonary perfusion images could be reduced in patients with left bundle block. Heights and weights of the patients were measured to calculate the body mass index. Heart rate, smoking status, as well as details on respiratory and cardiovascular diseases were collected from the medical records of the patients.

2.B | Study design and compliance with ethical guidelines
This was a retrospective case-control study investigating the feasibility of the breath-holding protocol in chest DDR. The research was conducted according to the principles of the Declaration of Helsinki.
Written informed consent was provided by all patients prior to their participation. This study has been reviewed based on the Clinical Trials Act and approved as a specified clinical trial.

2.C | Imaging protocol of chest DDR
Posteroanterior chest DDR was performed using a flat panel detector system (Konica Minolta, Inc., Tokyo, Japan) composed of a flat panel detector (AeroDR fine; Konica Minolta, Inc.) and a pulsed x-ray generator (RAD speed Pro; Shimadzu Corporation, Kyoto, Japan).
Each patient was scanned in the standing position and the pelvis was firmly fastened using a belt (Fig. 2). All patients were instructed regarding the two protocols: breath-holding protocol and deepbreathing protocol (Fig. 3). The breath-holding protocol requested patients to breathe with deep inspiration and hold their breath for  Color-mapping images were produced to visualize the pulmonary perfusion (Videos S1 and S2). This visualizing method is similar to that used by Tanaka et al., 8   2.E | Imaging success, correlation rate, and exposure to radiation In this study, imaging success was defined as right diaphragm excursion of ≤5 mm during breath-holding (2 s). This is because the size of each ROI in the lung fields was 5 × 5 mm. Diaphragm movement by ≥5 mm could cause respiratory artifacts. We applied the same exposure condition to both protocols. The entrance surface dose was measured as the exposure to radiation, which was calculated assuming a body thickness of 20 cm.

2.D | Image analysis
The correlation rate was calculated from the MIP images. We investigated the correlation rate between each pixel values of the lung fields and those of the left cardiac ventricles. The correlation rate calculation involves three steps. First, identification of the pulmonary field and its delimitation into 5-mm blocks. Second, extraction of the left ventricular signal change during one cardiac cycle.
Third, calculation of the percentage of pixels that match the left ventricular signal within these blocks. The lung fields of the MIP were equally divided into three parts from the top (upper lung field, middle lung field, and lower lung field).

2.F | Lung function tests
All patients underwent lung function tests by routine spirometry using Superior Spiro Discom-21 FXIII (Chest, Tokyo, Japan). The parameters of forced vital capacity (FVC), percent forced vital F I G . 2. Clinical practice of chest digital dynamic radiography (DDR). (a) Flat panel detector system (Konica Minolta, Inc., Tokyo, Japan) was composed of a flat panel detector (AeroDR fine; Konica Minolta, Inc., Tokyo, Japan) (white arrowhead) and a pulsed x-ray generator (RAD speed Pro; Shimadzu Corporation, Kyoto, Japan) (black arrowhead). In addition to chest DDR, conventional chest radiographs can also be taken. (b) Standing posteroanterior position in chest digital dynamic radiography. In order to minimize motion artifact, the pelvis is fastened firmly by a belt (white arrow).

3.A | Study population
Patient demographics and baseline characteristics are shown in Table 1. 3.B | Imaging success rate and exposure to radiation The imaging success rate was higher with the breath-holding protocol vs the deep-breathing protocol (97% vs 69%, respectively) (P < 0.0001). The entrance surface dose was lower with the breath-holding protocol (1.09 ± 0.20 vs 1.81 ± 0.08 mGy, respectively; P < 0.0001) ( Table 2). In the deep-breathing protocol, a univariate analysis of associations between excursion of the diaphragm and patient characteristics demonstrated that reduction in FEV 1 and %FEV 1 correlated with increase in excursion of the right diaphragm (Table 3). For the breath-holding protocol, there were no correlations between excursion of the diaphragm and patient characteristics.

3.C | Comparison of the correlation rate between the breath-holding protocol and deep-breathing protocol
The correlation rate was higher with the breath-holding protocol vs the deep-breathing protocol (Table 4). In the lower lung fields, the correlation rate was markedly different. The difference between the two protocols in the right and left lower lung fields was 15.3% and 14.1%, respectively (both P < 0.0001). Compared with each lung field, the correlation rate of the middle lung fields exhibited the highest value in both protocols.
Chest DDR images of a 74-yr-old male with chronic obstructive pulmonary disorder (COPD) are presented as an example. Color-mapping images of chest DDR conducted using both protocols were generated (Videos S1 and S2). The pulmonary perfusion of the bilateral lower lung fields became vague (blurred image) compared with that obtained using the breath-holding protocol. Horizontal lines observed in the lung field were regarded as incomplete pause of the diaphragm and ribs (misalignment of the rib cage). Similar findings were recorded using the MIP images [ Figs. 4(b) and 4(c)].   Values are presented as the mean ± SD unless otherwise indicated. All the results were analyzed using the paired t test. SD, standard deviation. *P < 0.05.
T A B L E 3 Univariate analysis of associations between excursion of the diaphragm and patient characteristics.
Breath-holding protocol for excursion of the right diaphragm Deep-breathing protocol for excursion of the right diaphragm FEV 1 and %FEV 1 . For the breath-holding protocol, there were no correlations between excursion of the diaphragm and patient characteristics. The reduction of respiratory artifacts noted after using the breath-holding protocol led to the capture of vivid pulmonary perfusion images, including high concordance with the left cardiac ventricle. In addition to the protocol, there are two other differences between the present study and previous studies. 11,12,15,16 The first is that the present study included the largest number of patients examined in this field. Thus far, only a few number of case reports, preliminary studies, and animal experiments 14 have been reported on DDR technology for pulmonary perfusion. The second is that our study clarified the real-world challenges of the deep-breathing protocol. Previous studies have shown that the deep-breathing protocol has the advantage of allowing the simultaneous evaluation of ventilation and blood flow. 8 However, in some patient groups (e.g., elderly patients or those with respiratory disease), the quality of the pulmonary perfusion images was lower than desired.
Incomplete diaphragm pause could be the main cause of respiratory artifacts. Bizarre diaphragm motion may occur among patients with various respiratory diseases (especially COPD). [17][18][19] Thus, the imaging protocol is required to reduce excursion of the diaphragm. The breath-holding protocol is able to assess pulmonary perfusion in all patients, even those with low FEV 1 /%FEV 1 . Furthermore, the shorter examination time resulted in less exposure to ionizing radiation than that linked to the deep-breathing protocol.
According to the study conducted by Tanaka  Values are presented as the mean ± SD unless otherwise indicated. All the results were analyzed using the paired t test. SD, standard deviation. *P < 0.05.
with the expansion of all lung lobes. 23 Thus, our breath-holding protocol has some advantages, namely the timely detection of pulmonary blood flow, shorter examination time, and exposure to lower levels of radiation.
There are three limitations in our study. First, this was a retrospective study conducted at a single center with a small number of patients, which limits the generalizability of the results. For example, there were only a few females, and patients with interstitial lung disease, bronchial asthma, or lung cancer in this study.
Therefore, a larger multicenter study is warranted. Second, many previous studies have used other modalities (contrast CT, magnetic resonance angiography, and scintigraphy) to evaluate pulmonary perfusion. Comparison with other modalities was beyond the scope of this study. Hence, further research to compare chest DDR with other modalities is required. Third, we investigated respiratory artifacts but not body motion artifacts. Color-mapping/MIP images may also contain body motion artifacts. In this study, we firmly fastened the patients' lumber with a belt to minimize the body motion artifacts.

| CONCLUSION
We have illustrated a new approach to image pulmonary perfusion by improving the protocol of chest DDR. For the evaluation of pulmonary perfusion, the breath-holding protocol was more beneficial than the deep-breathing protocol among patients with respiratory diseases, and resulted in a significant reduction of respiratory artifacts. Vivid images of pulmonary perfusion can be obtained through a simple and rapid examination. We are confident that our research will serve as a basis for future studies on various abnormalities of pulmonary perfusion.